For force and pressure sensing, the Columbia Team obtained a demo kit from Tekscan.
More info can be found at https://www.tekscan.com/product-group/embedded-sensing/force-sensors.
"These are single pad piezoresistive sensors a few mm in diameter ... with three wires trailing out of a mylar film ~10cm long or so. Two of these wires are used: they are passed through an op-amp to connect to an A/D converter. We'll be hooking them to an Arduino board, at first through whatever A/D converter is available.
• Greater flexibility
• Superior linearity & accuracy (±3%)
• Expert technical guidance
• Wide range of forces
• Sensor output is not a function of loading area
• High temperature force measurements (up to 400ºF) available with High-Temp Model (HT201)
• Optional sensing area sizes
• Off-the-shelf availability for testing & prototyping
Perhaps the most interesting point here is "[force measurement] not a function of load area", which needs to be true if force is to be measured not pressure. If we can get a sensor element small enough (see TekScan pressure sensors) then single grains 2-4mm across, and with contact areas much smaller than that, should provide us with direct measurement of roughly normal forces acting on the wall or bed (depending on where we locate them) of the drum (or linear heap once built). We can test this to some extent and will do.
In the near future we'd like to give a TekScan pressure array sensor a try - and this may be necessary simply to get a sensor elements small enough. What's not clear to me right now is whether these is any qualitative difference between these piezoresistive sensor elements and the single element in a FlexiForce sensor. Presumably not, and presumably "pressure" is only used for marketing purposes because the array of force measurements can be mapped into a continuous pressure field - the basis of TekScan's iScan system."
